The current ratings indicated on the schematic for supplies are the max. capabilities of the LDOs. The actual current draw per rail is significantly lower, e.g. for the analogue rail
Nice video Phil. I worked with AD for three years but decades with Cadence/Allegro and Pads. I was very happy to hear your philosophy of a single ground with the qualifier that you need to know what you're doing. One thing that I've found very helpful and used regularly on complex boards is the concept of rooms. Not only for the PCB designer, rooms remind you of where things are and not just components, but voltages, currents and their frequencies, and I/O. Fortunately, AD supports rooms, although not necessarily with the easiest implementation. And your lovely schematics, with sections identified and labeled, lend themselves directly to rooms. One message that pretty much rules in PCB design, echoed by the gurus like Hartly and Bogatin, is to know where your current flows, know where it returns, and to treat that path with respect. Decoupling capacitors bring the energy for the task to the local area where it's needed. Your high current, whether analog or digital, should have it's transient requirements satisfied by capacitors including bulk, "normal distributed decoupling" and for RF and certain compliance reasons, those caps which are right at the IC pins. Basic boards are important to get your blood flowing, cut your teeth, etc. But when your boards get more complex, contain mixed signals AND may need to pass compliance testing in production, the rules for signal flow and returns will never/seldom fail you. While AD is a great, powerful, professional packages, I have a voodoo doll and pins out for Altium for their abandonment of Circuit Studio and the evolution of it to CircuitMaker Pro, presumably sharing code base and many feature of the (especially now) very high priced package. There are many developers out there, and hobbyists who cannot afford the full AD seat. Students of course, can get the full package for (I guess) free or at least very cheap. And contractors of course, may also need to support other packages (Cadence/Allegro, Pads, etc.) without full installs. This ends my jab at Altium for abandoning the low end market. But I fully understand the business decisions regarding a $500 package with perpetual license and a $5000 package (just a 1 year license). While I hate subscriptions, I realize the company needs to survive.
Great comment! Regarding your last sentence: it's never about survival or steady state. Any corporation's fiduciary responsibility and actually its whole raison d'être, is to generate returns for investments. And if those returns fail to grow year over year, every year - the investors will take the moneys out and put them elsewhere, where they do grow. A corporate manager's job, if they know what's good for them and want that fat bonus, is not merely matching last year's growth figures - that are now Wall St.'s projections for this year, but to blow them out of the water. That pharma bro Martin Shkreli's only mistake was, he'd bit off too much at once for how poorly connected he was politically and how scrawny he looked optically - but otherwise this is what's done every day in every field. Everyone's riding their buggy till its wheels come off. Granted, there are CEOs and even entire companies that defy that picture. In Wall Street lingo, they're "leaving money on the table". And it's a matter of time before they're acquired & taken off the market by their peers that haven't been leaving any money on the table.
Thank you for posting this PCB design and recommendations/tips, Phil. It is very close to my current, first, and only SMD PCB in Kicad for an ILDA standard, 6 channel analog input -> TL084 op amp conversion circuit -> CS42448 ADC CODEC -> TDM -> Teensy 4.1 laser synth -> TDM -> CS42448 DAC -> 6 channels audio output -> TL084 op amp conversion circuits -> ILDA standard DB25..... PCB. I've assembled a working prototype, using a T4.1, an OTC CODEC PCB & my own op amp circuits on a breadboard. but, had the dreaded noisy power rails, of which I've mitigated with some large electrolytic caps. Full disclosure: I'm only a hobbyist and far from being an EE. (Can't believe how far I've gotten with this project, TBH.) In Kicad's PCB Editor, I've already placed all of the analog inputs on the left, the T4 & CODEC IC in the center, and all of the analog outputs on the right. So, with everything positioned, I've started laying tracks for a 2 layer PCB. But, within only 15 minutes, you've already pointed out several of my design/layout errors. Thank you for correcting me and avoiding my wasting $ on garbage. Try, try, again. 10th time lucky, I hope? Love your informative, clearly narrated channel... at least the parts that I can understand. Cheers, Phil!
Yes, please make a video for detailed description of the design. Especially about the analog input/output stages. And thanks for all the valuable content!
Hey Phil, great video as always! I, for one, would love to see a video focused on the design of this effects board! Also, I just wanted to say that I appreciate it when you get into tips regarding component selection. I find component selection hugely intimidating, so I always get a lot of value out of those sections. Keep up the good work!
Hey Louis, Thank you - I'm hopefully getting the enclosures for this board soon, then will make a more detailed video. I'll aim to make more videos on component selection as well!
Hey Phil, thanks for this great content. I also think the same way as you do when it comes to "not splitting ground planes". I learned this reality by watching videos from experts such as Rick Hartley and Eric Bogatin. But, recently I went back to my notes and recognized something. In a training video by Rick Hartley named "How to Achieve Proper Grounding - Rick Hartley - Expert Live Training (US)" at 1:13:30 it says that "... the need to split Analog & Digital Ground is only necessary when the analog section is operating under 20kHz". This is the case for an audio signal conditioning circuit. What are your thoughts on this? Do we need to split planes if we work with audio signals?
I'll add my $0.02 regarding this. It all depends on the type of analog signal. If you're doing A to D conversion, think of the number of bits you're converting. A signal being encoded to only 8 bits can live in a somewhat noisy environment. Audio encoded to 16 bits, says for a stereo music player, needs another level of caution and separation. By the time you get to 24 bits or more, say in a high end digital recording system, you will be highly isolating analog, digital and the analog side of the chips involved. Depending on the product size and other hardware, you may also be introducing metal shields over critical areas. Shields are to protect you from noise but also, in other scenarios, to contain the noise that you are generating in order to pass compliance tests.
I like the explanation of digital signals being separated from analog signals. In the Atmel reference for the ATMega32 for instance, it shows a dedicated analog ground plane for the ADC, but it doesn't really specify how this ground plane connects to the common ground plane. Is this different when considering ADC on the microcontroller? I am guessing from your tip, it is best to have the ground plane common but with stitching and traces connected to that portion of the ground plane as long as they are separated from the digital traces, it will be fine. This definitely makes this much more understandable. Thanks.
Thanks! Especially for MCUs, from my experience and from what I've gathered from others, it's pretty much never necessary (or a good idea) to split ground there. Some filtering (PI filter, beads + caps) on the MCU's analogue supplies + routing the analogue traces quickly away from any digital traces is what I've been doing and have had no issues so far. App notes/datasheets love recommended some rather 'odd' PCB design guidelines, as they're not typically written by PCB designers.
Ive seen both philosophies on ground plains, on the split between analog and digital. On those designs who decided to split the grounds, i have had more issues troubleshooting EMI issues than those who just combined everything.
FX1_EN_A from the footswitch to the cap directly can cause arcing on your footswitch. This can cause oxidation and premature aging of the contact. All such connections of switch to GND (or VCC) to a capacitor should have a series resistor to to current limit the in rush current. This in rush current can also exceed the peak current of the caps (2.2 ceramic would be ok, but tantalum or electrolytic are subject to maximum currents.) Without seeing the rest of the circuit, FX1_EN should be connected to a Schmitt trigger input because of the slow rise time created by the LP (pi-filter)
Hi Phil! Could you make a video about DSP theory with practical application exploring the feature that some STM32 have, the FMAC (filter math accelerator)? Thanks
Thanks again for the great content! I designed an audio board around the TLV320 codec and I had a question about that choice... Edit. You answered the question about the choice of codecs in the video. I removed it to avoid some personal embarrassment.
Great video! I have a question: If there are really huge space constraints, for example, a little usb 2x4cm aprox dongle with a microcontroller and a dac (4 layer: signal1/pwr - GND - GND - signal2/pwr)... Apart from separating the max possible in 2D and 3D (top digital / bottom analog), would you tie all ground vias across all the ground layers even being so close in space? For example, wouldn't be better to to put GND vias only for layer 1 to GND on layer 2, then vias for layer 4 to GND on layer 3 and then tie them together on a especific point of the pcb where there's nothing going on? (For example a corner, edges away from the action). Your videos are so informative dude, thanks for sharing your knowledge.
Yet another great video. Would love a video that goes a little into proper grounding on low frequency analog designs where there can be some use in not using a completely solid ground plane (at least if I have understood Rick Hartley correctly)
Thanks, Biko. That's a good idea! I wanted to make a test PCB for a future video with different grounding methods, looking at EMI and noise performance. Also including lower-frequency analogue (e.g. audio) grounding strategies.
Exactly when u want to design pcbs for biosignal acquisition, it is important to star connect the ground as the analog front end have sensitivities in terms of uV..
I'm starting to think I might have bitten of more than I can chew with my latest project 😅 Guess I'll have to see if there are any LSBs left to loose by my ADCs after I try my hand at this analog wizardry.
Hi Phil, Thanks again for sharing such useful tips and information. Always been big follower of your teachings. You have changed my perspective to look at electronic design, bringing excitement and motivation to follow designing as career. I have also decided to participate in your Altium design error finding contest, seeing your photo on it. Also eagerly waiting for your BGA course. Cutting my lunch cost and saving money. I don't want to miss that course. - KD Nayar
Thank you for your kind comment, KD. I'm glad to hear the content has been helpful and motivating - definitely much more to come! :) I hope you aren't cutting lunch money/costs too much, I'll be giving out discount coupons as well with the release. Thanks for your support!
Thank you Phil. I will reach out to you through mail, if I need any help with coupons. Currently I'm trying my best to save enough money. Also entered in many contests to earn prize money. Let's see. I really want to somehow enroll in your courses. You have taught me more than my 4 years of graduation, where my parents spend lots of money. Just for degree.
Great video! One comment; your DSP designs are generally not consuming and power in the analog section. Would be great to see a video focusing on a mixed signal design with higher current consumptions in the analog section - which would maybe need GND splitting.
Thanks for the tips Phil, really helpful. Quick question. You mentioned that the analog supplies should be assembled as far away from digital supplies as practicable. Do you have a comment about some of the PMUs available from Analog Devices and STM that integrate an LDO inside the same package as a number of buck regulators? e.g., ADP5050. Best regards.
Always a pleasure to pick up tips! Got any suggestions for reading matter regarding Audio and grounds ? as there are many opinions on the topic...cheers. And Merry Christmas to you and your's.
Thanks, Andy! Yes, I'm planning on making a video on specifically more analogue designs, including talking about grounding. Merry Christmas to you as well!
Very helpful video as always. I hope for a boost converter design video one day as it is a block that appears in almost every design and it has several complex and simple things altogether
Thanks to information and tips. And i am a little confused. What do you think about the motor driver? Should we separate the GNDs when designing an AC Induction motor driver? Thaks.
Thank you very much for this and all the videos. For real have been very teaching and significant. I'm now very exciting for the course you mentioned, I think will be amazing :) Thanks Phil.
Dear Phil, thank you for all the great content you provide on your youtube channel. I have a problem with my board and wanted to ask about your idea. I have an EG915 GPRS modem in my two-layer board and switching regulator (TPS54202) and an STM32 board. Considering that once EG915 wants to send the data it draws about 2A and the voltage signal gets distorted, so I decided to split the grounds. There is a ground for EG915 and TPS54202 and another ground for STM32. These grounds are connected with each other at one point, but there is a problem now, I got noise on my serial communication (UART) with the STM32 and the EG915. Do you have any suggestions on how to solve it?
Thank you so much for this I been trying to design this, but I'm self-taught and still learning. Also I'm designing a guitar power board 24v in and have it split into 9v 12v and 24v pass threw for guitar petals. I would love to see your idea or take on how to properly filter voltage so not to pick up on guitar amps.
Hey Phil; big fan of your designs and tutorials as a new engineer and electronics hobbyist. I play with some audio stuff, and I was wondering what your thoughts are on reference designs like the OPA1632D differential audio amp datasheet where it recommends removing the ground fills under the amplifier output/inverting pins? I've always followed these recommendations as I don't see the harm when dealing with low-frequency signals, and I didn't want any capacitively coupled gremlins to invade my differential opamp circuitry, but I'm wondering if this approach presents more good or bad.
Hello Phil, just wanted to thank you so much for your videos. I am an Elektrotechniker and now has an ingeneers job and doing hardware development. Your videos are so helpfully to me. Just wanted to ask, if you maybe can add an LDO on any of your PCBs and try to explain what is necessary for this. Often i have problems with the right capacitor values and also choosing a capacitor with the right ESR and so on. Right now i have an LDO from 6V to 3.3V and if i am using only ceramic capacitors, i have a big ripple. If i use the same vale as an electolytic capacitor, there is no ripple anymore. So i think it is because of the ESR. Maybe in future, you can try to explain this on another project and so on :) Would be helpfull. But right now, i hope you have a nice christmas and you will have a nice new year! Hope to see many new videos from you, next year :) Have a nice time!
Moin Thomas, Thank you - glad to hear the videos are helpful! Regarding LDOs, all information should be available in the respective datasheet. Also what caps to use and if a min. ESR rating is needed.
Thank you for the video, I have a problem with a mm phono preamp I made a pcb for, it works and sounds good with no audible noise when I have one channel hooked up but when I hook up the second channel which BTW is on a separate pcb I get a low frequency oscillation when turned up a little with the bass turned up a bit as well on the preamp. What has me puzzled is it has no problem when one of the rca phono channels is hooked up. Any insight would be great, thanks.
Hi Phil, Great video as always. I’m wanted to getting deep into EMI/SI/noise and was wondering what’s the detail process on designing a Pi filter for the application? Thanks!
Hey, Thanks! I'd highly recommend the book "EMC for Printed Circuit Boards" by Keith Armstrong. Keith also has a website with a lot of free material. I'm planning on making a ferrite bead (which will then cover pi filters) video but that may still be a few weeks.
Hi Phil i wanna asking one thing i designed a PCB with Analog IC (ADS1220IPWR) or DAC (DAC161S997RGHT) i gave separate power both analog and digital all track are separate but now the problem is both ic do not work together one ic remove than other ic work this circuit is for 4 to 20 m Amp. but both is separate on separate 2 pcb different different that they are working what is the region behind it.
Hello Phil, I LOVE your videos, I am a parttime PCB designer myself and have learnt a lot from you over the years. I have used altium for personal use but I can not figure out an easy way to make the edge plating or exposed copper just like your pcb around the edges. How is that done in altium? Thanks :)
Flatter surfaces for soldering, longer shelf life, better looks - to name a few reasons. Exposed GND copper is very nice as a very accessible test point for oscilloscope probes, for example.
@@PhilsLab This would be very interesting but I meant a proper course on a platform like Zoom where we register and get to interact directly with you and ask questions. I was thinking a live version of your mixed signal or upcoming FPGA course.
Hey, Thanks for watching! Click on a net in the schematic editor, then in the Properties panel you can assign a colour. Toggle colouring by pressing F5.
Do u keep the width between traces and the ground plane larger because interference is introduced by the ground plane? My question, why keep traces so much spaced apart from the ground plane.
The current ratings indicated on the schematic for supplies are the max. capabilities of the LDOs. The actual current draw per rail is significantly lower, e.g. for the analogue rail
Nice video Phil. I worked with AD for three years but decades with Cadence/Allegro and Pads. I was very happy to hear your philosophy of a single ground with the qualifier that you need to know what you're doing. One thing that I've found very helpful and used regularly on complex boards is the concept of rooms. Not only for the PCB designer, rooms remind you of where things are and not just components, but voltages, currents and their frequencies, and I/O. Fortunately, AD supports rooms, although not necessarily with the easiest implementation. And your lovely schematics, with sections identified and labeled, lend themselves directly to rooms.
One message that pretty much rules in PCB design, echoed by the gurus like Hartly and Bogatin, is to know where your current flows, know where it returns, and to treat that path with respect. Decoupling capacitors bring the energy for the task to the local area where it's needed. Your high current, whether analog or digital, should have it's transient requirements satisfied by capacitors including bulk, "normal distributed decoupling" and for RF and certain compliance reasons, those caps which are right at the IC pins. Basic boards are important to get your blood flowing, cut your teeth, etc. But when your boards get more complex, contain mixed signals AND may need to pass compliance testing in production, the rules for signal flow and returns will never/seldom fail you.
While AD is a great, powerful, professional packages, I have a voodoo doll and pins out for Altium for their abandonment of Circuit Studio and the evolution of it to CircuitMaker Pro, presumably sharing code base and many feature of the (especially now) very high priced package. There are many developers out there, and hobbyists who cannot afford the full AD seat. Students of course, can get the full package for (I guess) free or at least very cheap. And contractors of course, may also need to support other packages (Cadence/Allegro, Pads, etc.) without full installs.
This ends my jab at Altium for abandoning the low end market. But I fully understand the business decisions regarding a $500 package with perpetual license and a $5000 package (just a 1 year license). While I hate subscriptions, I realize the company needs to survive.
Great comment! Regarding your last sentence: it's never about survival or steady state. Any corporation's fiduciary responsibility and actually its whole raison d'être, is to generate returns for investments. And if those returns fail to grow year over year, every year - the investors will take the moneys out and put them elsewhere, where they do grow. A corporate manager's job, if they know what's good for them and want that fat bonus, is not merely matching last year's growth figures - that are now Wall St.'s projections for this year, but to blow them out of the water. That pharma bro Martin Shkreli's only mistake was, he'd bit off too much at once for how poorly connected he was politically and how scrawny he looked optically - but otherwise this is what's done every day in every field. Everyone's riding their buggy till its wheels come off.
Granted, there are CEOs and even entire companies that defy that picture. In Wall Street lingo, they're "leaving money on the table". And it's a matter of time before they're acquired & taken off the market by their peers that haven't been leaving any money on the table.
Thank you for posting this PCB design and recommendations/tips, Phil.
It is very close to my current, first, and only SMD PCB in Kicad for an ILDA standard, 6 channel analog input -> TL084 op amp conversion circuit -> CS42448 ADC CODEC -> TDM -> Teensy 4.1 laser synth -> TDM -> CS42448 DAC -> 6 channels audio output -> TL084 op amp conversion circuits -> ILDA standard DB25..... PCB.
I've assembled a working prototype, using a T4.1, an OTC CODEC PCB & my own op amp circuits on a breadboard. but, had the dreaded noisy power rails, of which I've mitigated with some large electrolytic caps.
Full disclosure: I'm only a hobbyist and far from being an EE. (Can't believe how far I've gotten with this project, TBH.)
In Kicad's PCB Editor, I've already placed all of the analog inputs on the left, the T4 & CODEC IC in the center, and all of the analog outputs on the right. So, with everything positioned, I've started laying tracks for a 2 layer PCB.
But, within only 15 minutes, you've already pointed out several of my design/layout errors. Thank you for correcting me and avoiding my wasting $ on garbage.
Try, try, again. 10th time lucky, I hope?
Love your informative, clearly narrated channel... at least the parts that I can understand.
Cheers, Phil!
Phil, your work is just incredible! thanks for sharing these tips.
Thank you very much, Yaghiyah!
Yes, please make a video for detailed description of the design. Especially about the analog input/output stages. And thanks for all the valuable content!
Hey Phil, great video as always! I, for one, would love to see a video focused on the design of this effects board!
Also, I just wanted to say that I appreciate it when you get into tips regarding component selection. I find component selection hugely intimidating, so I always get a lot of value out of those sections. Keep up the good work!
Hey Louis, Thank you - I'm hopefully getting the enclosures for this board soon, then will make a more detailed video.
I'll aim to make more videos on component selection as well!
Hey Phil, thanks for this great content. I also think the same way as you do when it comes to "not splitting ground planes". I learned this reality by watching videos from experts such as Rick Hartley and Eric Bogatin. But, recently I went back to my notes and recognized something. In a training video by Rick Hartley named "How to Achieve Proper Grounding - Rick Hartley - Expert Live Training (US)" at 1:13:30 it says that "... the need to split Analog & Digital Ground is only necessary when the analog section is operating under 20kHz". This is the case for an audio signal conditioning circuit. What are your thoughts on this? Do we need to split planes if we work with audio signals?
I'll add my $0.02 regarding this. It all depends on the type of analog signal. If you're doing A to D conversion, think of the number of bits you're converting. A signal being encoded to only 8 bits can live in a somewhat noisy environment. Audio encoded to 16 bits, says for a stereo music player, needs another level of caution and separation. By the time you get to 24 bits or more, say in a high end digital recording system, you will be highly isolating analog, digital and the analog side of the chips involved. Depending on the product size and other hardware, you may also be introducing metal shields over critical areas. Shields are to protect you from noise but also, in other scenarios, to contain the noise that you are generating in order to pass compliance tests.
Thank you again for an awesome video!
Thanks, Alexandros!
I like the explanation of digital signals being separated from analog signals. In the Atmel reference for the ATMega32 for instance, it shows a dedicated analog ground plane for the ADC, but it doesn't really specify how this ground plane connects to the common ground plane. Is this different when considering ADC on the microcontroller? I am guessing from your tip, it is best to have the ground plane common but with stitching and traces connected to that portion of the ground plane as long as they are separated from the digital traces, it will be fine. This definitely makes this much more understandable. Thanks.
Thanks! Especially for MCUs, from my experience and from what I've gathered from others, it's pretty much never necessary (or a good idea) to split ground there. Some filtering (PI filter, beads + caps) on the MCU's analogue supplies + routing the analogue traces quickly away from any digital traces is what I've been doing and have had no issues so far.
App notes/datasheets love recommended some rather 'odd' PCB design guidelines, as they're not typically written by PCB designers.
Ive seen both philosophies on ground plains, on the split between analog and digital. On those designs who decided to split the grounds, i have had more issues troubleshooting EMI issues than those who just combined everything.
FX1_EN_A from the footswitch to the cap directly can cause arcing on your footswitch. This can cause oxidation and premature aging of the contact. All such connections of switch to GND (or VCC) to a capacitor should have a series resistor to to current limit the in rush current. This in rush current can also exceed the peak current of the caps (2.2 ceramic would be ok, but tantalum or electrolytic are subject to maximum currents.) Without seeing the rest of the circuit, FX1_EN should be connected to a Schmitt trigger input because of the slow rise time created by the LP (pi-filter)
My favourite channel
Thank you, glad to hear that :)
Hi Phil,
If you can, please do some in depth video of this board! It would help us so much!
Awesome video with great content and a good cadence as always, Phil 👍 BTW, have a wonderful festive season! ☃⛄🎄🎄
Thank you very much 😊 I hope you all at PCBWay have a good start to the new year!
Hi Phil! Could you make a video about DSP theory with practical application exploring the feature that some STM32 have, the FMAC (filter math accelerator)? Thanks
Thanks again for the great content! I designed an audio board around the TLV320 codec and I had a question about that choice...
Edit. You answered the question about the choice of codecs in the video. I removed it to avoid some personal embarrassment.
I think I saw your comment before you edited it - glad the answer was in the video :)
What are the two slots used for? Why did you put them? Don't they break the ground plane? Thank you, Phil.
Great video!
I have a question: If there are really huge space constraints, for example, a little usb 2x4cm aprox dongle with a microcontroller and a dac (4 layer: signal1/pwr - GND - GND - signal2/pwr)... Apart from separating the max possible in 2D and 3D (top digital / bottom analog), would you tie all ground vias across all the ground layers even being so close in space? For example, wouldn't be better to to put GND vias only for layer 1 to GND on layer 2, then vias for layer 4 to GND on layer 3 and then tie them together on a especific point of the pcb where there's nothing going on? (For example a corner, edges away from the action).
Your videos are so informative dude, thanks for sharing your knowledge.
I do mostly RF layout, also use Altium. all manual layout, no automated routing.
Yet another great video. Would love a video that goes a little into proper grounding on low frequency analog designs where there can be some use in not using a completely solid ground plane (at least if I have understood Rick Hartley correctly)
Thanks, Biko. That's a good idea! I wanted to make a test PCB for a future video with different grounding methods, looking at EMI and noise performance. Also including lower-frequency analogue (e.g. audio) grounding strategies.
Exactly when u want to design pcbs for biosignal acquisition, it is important to star connect the ground as the analog front end have sensitivities in terms of uV..
I'm starting to think I might have bitten of more than I can chew with my latest project 😅
Guess I'll have to see if there are any LSBs left to loose by my ADCs after I try my hand at this analog wizardry.
Hi Phil,
Thanks again for sharing such useful tips and information. Always been big follower of your teachings.
You have changed my perspective to look at electronic design, bringing excitement and motivation to follow designing as career. I have also decided to participate in your Altium design error finding contest, seeing your photo on it.
Also eagerly waiting for your BGA course. Cutting my lunch cost and saving money. I don't want to miss that course.
- KD Nayar
Thank you for your kind comment, KD. I'm glad to hear the content has been helpful and motivating - definitely much more to come! :)
I hope you aren't cutting lunch money/costs too much, I'll be giving out discount coupons as well with the release. Thanks for your support!
Thank you Phil. I will reach out to you through mail, if I need any help with coupons. Currently I'm trying my best to save enough money. Also entered in many contests to earn prize money. Let's see. I really want to somehow enroll in your courses. You have taught me more than my 4 years of graduation, where my parents spend lots of money. Just for degree.
Great video! One comment; your DSP designs are generally not consuming and power in the analog section.
Would be great to see a video focusing on a mixed signal design with higher current consumptions in the analog section - which would maybe need GND splitting.
Thanks for the tips Phil, really helpful.
Quick question. You mentioned that the analog supplies should be assembled as far away from digital supplies as practicable. Do you have a comment about some of the PMUs available from Analog Devices and STM that integrate an LDO inside the same package as a number of buck regulators? e.g., ADP5050. Best regards.
Always a pleasure to pick up tips! Got any suggestions for reading matter regarding Audio and grounds ? as there are many opinions on the topic...cheers. And Merry Christmas to you and your's.
Thanks, Andy! Yes, I'm planning on making a video on specifically more analogue designs, including talking about grounding. Merry Christmas to you as well!
Please go into more detail for this PCB design in the future
Will do!
Very helpful video as always. I hope for a boost converter design video one day as it is a block that appears in almost every design and it has several complex and simple things altogether
Thanks! Will definitely cover that in the future.
Really cool! What about a high power (50Vdc 75A) together with a microcontroller circuit?
Thanks to information and tips. And i am a little confused. What do you think about the motor driver? Should we separate the GNDs when designing an AC Induction motor driver? Thaks.
Thank you very much for this and all the videos. For real have been very teaching and significant. I'm now very exciting for the course you mentioned, I think will be amazing :) Thanks Phil.
Thanks a lot for your support, Ricardo! :)
Another great video, so to speak.
Thanks!
Dear Phil, thank you for all the great content you provide on your youtube channel. I have a problem with my board and wanted to ask about your idea. I have an EG915 GPRS modem in my two-layer board and switching regulator (TPS54202) and an STM32 board. Considering that once EG915 wants to send the data it draws about 2A and the voltage signal gets distorted, so I decided to split the grounds. There is a ground for EG915 and TPS54202 and another ground for STM32. These grounds are connected with each other at one point, but there is a problem now, I got noise on my serial communication (UART) with the STM32 and the EG915. Do you have any suggestions on how to solve it?
Awesome video ,thanks for sharing.
Thanks, Annacer!
I love your videos. Thank you for sharing your great work!
Thanks for watching!
Thank you so much for this I been trying to design this, but I'm self-taught and still learning. Also I'm designing a guitar power board 24v in and have it split into 9v 12v and 24v pass threw for guitar petals. I would love to see your idea or take on how to properly filter voltage so not to pick up on guitar amps.
Thanks, Jason - good luck with your new design! I'm planning on making a video on power supply filtering in the near future.
I am still getting my head around 4 layer boards, its going to be a while before i get to mixed signal but still a great episode.
Thanks, Ian!
Hey Phil; big fan of your designs and tutorials as a new engineer and electronics hobbyist. I play with some audio stuff, and I was wondering what your thoughts are on reference designs like the OPA1632D differential audio amp datasheet where it recommends removing the ground fills under the amplifier output/inverting pins? I've always followed these recommendations as I don't see the harm when dealing with low-frequency signals, and I didn't want any capacitively coupled gremlins to invade my differential opamp circuitry, but I'm wondering if this approach presents more good or bad.
Are red polygons on the top layers ground too?
What is purpose of the gold color outline of the PCB and the vias in it?
Hello Phil,
just wanted to thank you so much for your videos. I am an Elektrotechniker and now has an ingeneers job and doing hardware development.
Your videos are so helpfully to me. Just wanted to ask, if you maybe can add an LDO on any of your PCBs and try to explain what is necessary for this.
Often i have problems with the right capacitor values and also choosing a capacitor with the right ESR and so on.
Right now i have an LDO from 6V to 3.3V and if i am using only ceramic capacitors, i have a big ripple. If i use the same vale as an electolytic capacitor, there is no ripple anymore. So i think it is because of the ESR. Maybe in future, you can try to explain this on another project and so on :) Would be helpfull.
But right now, i hope you have a nice christmas and you will have a nice new year! Hope to see many new videos from you, next year :)
Have a nice time!
Moin Thomas, Thank you - glad to hear the videos are helpful!
Regarding LDOs, all information should be available in the respective datasheet. Also what caps to use and if a min. ESR rating is needed.
Thank you for the video, I have a problem with a mm phono preamp I made a pcb for, it works and sounds good with no audible noise when I have one channel hooked up but when I hook up the second channel which BTW is on a separate pcb I get a low frequency oscillation when turned up a little with the bass turned up a bit as well on the preamp. What has me puzzled is it has no problem when one of the rca phono channels is hooked up. Any insight would be great, thanks.
Hi Phil,
Great video as always. I’m wanted to getting deep into EMI/SI/noise and was wondering what’s the detail process on designing a Pi filter for the application?
Thanks!
Hey, Thanks! I'd highly recommend the book "EMC for Printed Circuit Boards" by Keith Armstrong. Keith also has a website with a lot of free material. I'm planning on making a ferrite bead (which will then cover pi filters) video but that may still be a few weeks.
Phil, if you could find time to go in more detail on this design it'd be great. cheers.
Hi Phil i wanna asking one thing i designed a PCB with Analog IC (ADS1220IPWR) or DAC (DAC161S997RGHT) i gave separate power both analog and digital all track are separate but now the problem is both ic do not work together one ic remove than other ic work this circuit is for 4 to 20 m Amp. but both is separate on separate 2 pcb different different that they are working what is the region behind it.
Could the middle ground planes be just 1 so the board is just 3 layers?
Hello Phil, I LOVE your videos, I am a parttime PCB designer myself and have learnt a lot from you over the years. I have used altium for personal use but I can not figure out an easy way to make the edge plating or exposed copper just like your pcb around the edges. How is that done in altium? Thanks :)
Hi Milan, Thank you! I did it a rather manual way, by drawing traces/polygons and then copying that over to the soldermask layers.
Are you gonna make a discount for university students ? Because the price is very expensive for a student living in Egypt like me .
Why do you prefer inmersion gold? And, why do you use the exposed gold border? Thanks for your awesome videos.
Flatter surfaces for soldering, longer shelf life, better looks - to name a few reasons.
Exposed GND copper is very nice as a very accessible test point for oscilloscope probes, for example.
Phil, are you planning to cover design for CAN protocol in a future video?
At some point I'd definitely like to!
Thank you Phil for sharing these tips. Would you consider doing live courses in the future?
Thanks for watching, Khadem! As in a live stream where I design a board?
@@PhilsLab PCB design Live Stream?? We can ask for that? I am all for it.
@@PhilsLab This would be very interesting but I meant a proper course on a platform like Zoom where we register and get to interact directly with you and ask questions. I was thinking a live version of your mixed signal or upcoming FPGA course.
@@isaacclark9825 +1 I really like the idea of live PCB design.
Wow thanks. Do you have planed to do a video about Reverb Programming?
Yes, I've made a reverb algorithm that I'll be sharing as well!
Very good input! Thank you for this great content! May I ask which Codec you have used?
Thanks, Timo! I used a Cirrus Logic CS4270.
@@PhilsLab Thank you so much!
Hey Phil,
Thank you very much for your video. How are you changing the color of the trace? for example, at 7:12.
Thank you very much :)
Hey, Thanks for watching! Click on a net in the schematic editor, then in the Properties panel you can assign a colour. Toggle colouring by pressing F5.
Do you recall which audio codec you used in this design? Was it something like a PCM5102?
CS4270-CZZR - which is now unfortunately EOL.
@@PhilsLab No worries, thank you so much for the information!
Does the schematic file for this circuit avaialble open source
Do u keep the width between traces and the ground plane larger because interference is introduced by the ground plane? My question, why keep traces so much spaced apart from the ground plane.
Because analogue system does not effected by magnetic field
Can you be a guide to the power supply? plz😅😅
Thank you so much !!!❤❤❤
Thanks for watching!
Only time i ever split grounds is when i use canbus, as the ground of CANBUS is supposed to be connected to earth.
problem i have with altium is it dont work on mac or linux
Nice one.
Thanks!
good
Hey Phil, I'd love to see your videos on Odysee!
👍🙏❤
Comic Sans? Really?
Sorry :( At least it stands out :D